Dimension Error in train.py

[shiva811]

I am running train.py with the correct options and it is givinig me this error
21-05-24 21:22:32.620 - INFO: Loading model for G [/data2/superresolution/real-world-sr/esrgan-fs/experiments/Multiview_SR/models/55000_G.pth] ... 21-05-24 21:22:32.946 - INFO: Model [SRGANModel] is created. 21-05-24 21:22:32.946 - INFO: Start training from epoch: 0, iter: 0 /data2/superresolution/BasicSR/venv/lib/python3.8/site-packages/torch/nn/modules/loss.py:94: UserWarning: Using a target size (torch.Size([16, 3, 128, 128])) that is different to the input size (torch.Size([16, 3, 256, 256])). This will likely lead to incorrect results due to broadcasting. Please ensure they have the same size. return F.l1_loss(input, target, reduction=self.reduction) Traceback (most recent call last): File "train.py", line 244, in <module> main() File "train.py", line 163, in main model.optimize_parameters(current_step) File "/data3/shivam/lsdr/Real-SR/codes/models/SRGAN_model.py", line 147, in optimize_parameters l_g_pix = self.l_pix_w * self.cri_pix(self.fake_H, self.var_H) File "/data2/superresolution/BasicSR/venv/lib/python3.8/site-packages/torch/nn/modules/module.py", line 727, in _call_impl result = self.forward(*input, **kwargs) File "/data2/superresolution/BasicSR/venv/lib/python3.8/site-packages/torch/nn/modules/loss.py", line 94, in forward return F.l1_loss(input, target, reduction=self.reduction) File "/data2/superresolution/BasicSR/venv/lib/python3.8/site-packages/torch/nn/functional.py", line 2633, in l1_loss expanded_input, expanded_target = torch.broadcast_tensors(input, target) File "/data2/superresolution/BasicSR/venv/lib/python3.8/site-packages/torch/functional.py", line 71, in broadcast_tensors return _VF.broadcast_tensors(tensors) # type: ignore RuntimeError: The size of tensor a (256) must match the size of tensor b (128) at non-singleton dimension 3

Please can somebody help me resolve this error? I am looking at all the possible cases which could have caused this but couldn't find one yet.

I would be very thankful if you could offer some suggestions or advice.